SmartShift: Expanded Load Shifting Incentive Mechanism for Risk-Averse Consumers

Authors: Bochao Shen, Balakrishnan Narayanaswamy, Ravi Sundaram

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental demonstration of the practical benefits of Smart Shift through extensive simulations (Simulations section).We evaluate Smart Shift using a real-time electricity load data set from Smart* (Barker et al. 2012). Through numerical experiments, we then study how our expanded load shifting mechanism can improve the performance over flat-rate pricing while providing a win-win solution to both the distribution company as well as the consumer.
Researcher Affiliation Academia Bochao Shen CCIS Northeastern University Boston MA, 02115 ordinary@ccs.neu.edu Balakrishnan Narayanaswamy Computer Science and Engineering University of California, San Diego La Jolla CA, 92093 muralib@cs.ucsd.edu Ravi Sundaram CCIS Northeastern University Boston MA, 02115 koods@ccs.neu.edu
Pseudocode Yes Algorithm 2 Optimal load shifting under exogenous market price Input: {x(t) i |1 i n, 1 t k}, {m(s t) i |1 i n, 1 s k, 1 t k}. Output: {z(t) i |1 i n, 1 t k}. 1: for each consumer i do 2: for each time slot t do 3: z(t) i arg min 1 t k {m(t t ) i p(t ) m }; 4: end for 5: end for
Open Source Code No No explicit statement or link indicating the availability of open-source code for the described methodology was found.
Open Datasets Yes We evaluate Smart Shift using a real-time electricity load data set from Smart* (Barker et al. 2012).
Dataset Splits No The paper describes data processing and filtering ('average the (per minute) sampled data points per hour', 'filter out households with zero power usage'), but does not provide explicit training/validation/test dataset splits, percentages, or absolute sample counts.
Hardware Specification No No specific hardware details (like GPU/CPU models, memory, or cloud instance types) used for running experiments were mentioned in the paper.
Software Dependencies No The paper mentions simulating market prices using a normal distribution, but does not list any specific software components or libraries with version numbers (e.g., Python 3.8, PyTorch 1.9, CPLEX 12.4).
Experiment Setup Yes In all our simulations, each point is averaged over 100 repeated experiments. 95% confidence interval is also shown. In Figure 1, we draw the risk-aversion parameter {λi} from the Pareto distribution, Eq. (2), by setting βR = 1 and λmin = 1. These are fixed for subsequent repetitions. ... In Figure 2, we fix the fluctuation of the price by setting µm = 50, σm = 5, but vary the tolerance of the consumers by varying the βG from 0.5 to 5 with step size of 0.5. For each βG, we sample a set of prices from normal distribution N(µm, σ2 m) where µ = 50, σm = 5, then sample the {m(s t) i } from Eq. (3) with the fixed {m(s t) i,min} and the varying βG.